Motivation

Plastic pollution is a major and growing problem, negatively affecting oceans and wildlife health. Our World in Data has a lot of great data at various levels including globally, per country, and over time. This report will explore the distributions of plastic waste and mismanaged plastic waste generated per capita per day globally using data from 2016 obtained from a recent article. The data contain information on two different types of plastic waste generated by the populations in 211 countries as well as other variables such as income status, region, general population size,urban and coastal population size, as well as GDP value per capita.

Loading Data

load("~/Downloads/DS Durham/Data Visualization/Datasets/plastic.Rda")
head(plastic)

##            Country Code                     Region        GDP IncomeStatus
## AFG    Afghanistan  AFG                 South Asia   509.2187          LIC
## ALB        Albania  ALB      Europe & Central Asia  4124.0557          UMC
## DZA        Algeria  DZA Middle East & North Africa  3945.4821          UMC
## ASM American Samoa  ASM        East Asia & Pacific 11696.9556          UMC
## AND        Andorra  AND      Europe & Central Asia 37474.6654          HIC
## AGO         Angola  AGO         Sub-Saharan Africa  3506.0729          LMC
##     Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## AFG   35383128         0.00      25.02        11.45           NA
## ALB    2876101         7.07      58.42        39.32           NA
## DZA   40551404         0.76      71.46        51.59         1.03
## ASM      55741         3.27      87.20        43.61           NA
## AND      77297         0.00      88.25        63.73        31.80
## AGO   28842484         1.44      64.15        22.60           NA

names(plastic)

##  [1] "Country"      "Code"         "Region"       "GDP"          "IncomeStatus"
##  [6] "Population"   "CoastalPopPC" "UrbanPopPC"   "PlasticWaste" "MismanagedPW"

str(plastic)

## 'data.frame':    211 obs. of  10 variables:
##  $ Country     : chr  "Afghanistan" "Albania" "Algeria" "American Samoa" ...
##  $ Code        : Factor w/ 211 levels "ABW","AFG","AGO",..: 2 4 54 9 5 3 10 7 8 1 ...
##  $ Region      : Factor w/ 7 levels "East Asia & Pacific",..: 6 2 4 1 2 7 3 3 2 3 ...
##  $ GDP         : num  509 4124 3945 11697 37475 ...
##  $ IncomeStatus: Factor w/ 4 levels "LIC","LMC","UMC",..: 1 3 3 3 4 2 4 4 3 4 ...
##  $ Population  : num  35383128 2876101 40551404 55741 77297 ...
##  $ CoastalPopPC: num  0 7.07 0.76 3.27 0 ...
##  $ UrbanPopPC  : num  25 58.4 71.5 87.2 88.2 ...
##  $ PlasticWaste: num  11.4 39.3 51.6 43.6 63.7 ...
##  $ MismanagedPW: num  NA NA 1.03 NA 31.8 ...

# rename income status level for an easier analysis
levels(plastic$IncomeStatus) <- c("Low", "Low-middle", "High-middle", "High")

Data Exploration

# install.packages('skimr')
library(skimr)
skim(plastic)

Data summary
Name	plastic
Number of rows	211
Number of columns	10
_______________________
Column type frequency:
character	1
factor	3
numeric	6
________________________
Group variables	None

Variable type: character

skim_variable	n_missing	complete_rate	min	max	empty	n_unique	whitespace
Country	0	1	4	30	0	211	0

Variable type: factor

skim_variable	complete_rate	ordered	n_unique	top_counts
Code	1	FALSE	211	ABW: 1, AFG: 1, AGO: 1, ALB: 1
Region	1	FALSE	7	Eur: 55, Sub: 47, Lat: 42, Eas: 36
IncomeStatus	1	FALSE	4	Hig: 78, Hig: 56, Low: 44, Low: 33

Variable type: numeric

skim_variable	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
GDP	0	1.00	16496.91	25010.08	0.00	2053.79	5870.78	19717.33	1.700287e+05	▇▁▁▁▁
Population	0	1.00	34921473.92	136312697.76	11225.00	833995.50	6711121.00	23874626.50	1.378665e+09	▇▁▁▁▁
CoastalPopPC	0	1.00	6.10	10.40	0.00	0.33	2.50	6.28	5.851000e+01	▇▁▁▁▁
UrbanPopPC	0	1.00	59.60	24.68	0.00	40.84	60.18	79.71	1.000000e+02	▂▅▇▇▇
PlasticWaste	2	0.99	46.00	44.30	1.49	15.67	31.60	58.66	3.082500e+02	▇▂▁▁▁
MismanagedPW	41	0.81	17.42	25.66	0.14	3.13	10.09	19.39	1.430000e+02	▇▁▁▁▁

Missing Data Analysis

The data has one pattern of missingness in the mismanaged plastic waste variable, and two patterns of missingness in both types of plastic waste. Notably, there are 41 missing cases in the mismanaged plastic waste alone and two missing cases in the plastic waste. Around 56% of the total missingness in the mismanaged plastic waste is attributed to Sub-Saharan Africa.

# install.packages('mice', 'VIM')
library(mice)

## 
## Attaching package: 'mice'

## The following object is masked from 'package:stats':
## 
##     filter

## The following objects are masked from 'package:base':
## 
##     cbind, rbind

library(VIM)

## Loading required package: colorspace

## Loading required package: grid

## VIM is ready to use.

## Suggestions and bug-reports can be submitted at: https://github.com/statistikat/VIM/issues

## 
## Attaching package: 'VIM'

## The following object is masked from 'package:datasets':
## 
##     sleep

## first row: 170 complete cases
## second row:there are 39 patterns with only 1 missing value, which is mismanaged plastic waste
## third row: there are 2 patterns with 2 missing values, which are plastic waste and mismanaged plastic waste
## last row: there are 2 missing values in plastic waste and 41 missing values in mismanaged plastic waste
## overall, 43 observations are missing out of 211.

md.pairs(plastic)

## $rr
##              Country Code Region GDP IncomeStatus Population CoastalPopPC
## Country          211  211    211 211          211        211          211
## Code             211  211    211 211          211        211          211
## Region           211  211    211 211          211        211          211
## GDP              211  211    211 211          211        211          211
## IncomeStatus     211  211    211 211          211        211          211
## Population       211  211    211 211          211        211          211
## CoastalPopPC     211  211    211 211          211        211          211
## UrbanPopPC       211  211    211 211          211        211          211
## PlasticWaste     209  209    209 209          209        209          209
## MismanagedPW     170  170    170 170          170        170          170
##              UrbanPopPC PlasticWaste MismanagedPW
## Country             211          209          170
## Code                211          209          170
## Region              211          209          170
## GDP                 211          209          170
## IncomeStatus        211          209          170
## Population          211          209          170
## CoastalPopPC        211          209          170
## UrbanPopPC          211          209          170
## PlasticWaste        209          209          170
## MismanagedPW        170          170          170
## 
## $rm
##              Country Code Region GDP IncomeStatus Population CoastalPopPC
## Country            0    0      0   0            0          0            0
## Code               0    0      0   0            0          0            0
## Region             0    0      0   0            0          0            0
## GDP                0    0      0   0            0          0            0
## IncomeStatus       0    0      0   0            0          0            0
## Population         0    0      0   0            0          0            0
## CoastalPopPC       0    0      0   0            0          0            0
## UrbanPopPC         0    0      0   0            0          0            0
## PlasticWaste       0    0      0   0            0          0            0
## MismanagedPW       0    0      0   0            0          0            0
##              UrbanPopPC PlasticWaste MismanagedPW
## Country               0            2           41
## Code                  0            2           41
## Region                0            2           41
## GDP                   0            2           41
## IncomeStatus          0            2           41
## Population            0            2           41
## CoastalPopPC          0            2           41
## UrbanPopPC            0            2           41
## PlasticWaste          0            0           39
## MismanagedPW          0            0            0
## 
## $mr
##              Country Code Region GDP IncomeStatus Population CoastalPopPC
## Country            0    0      0   0            0          0            0
## Code               0    0      0   0            0          0            0
## Region             0    0      0   0            0          0            0
## GDP                0    0      0   0            0          0            0
## IncomeStatus       0    0      0   0            0          0            0
## Population         0    0      0   0            0          0            0
## CoastalPopPC       0    0      0   0            0          0            0
## UrbanPopPC         0    0      0   0            0          0            0
## PlasticWaste       2    2      2   2            2          2            2
## MismanagedPW      41   41     41  41           41         41           41
##              UrbanPopPC PlasticWaste MismanagedPW
## Country               0            0            0
## Code                  0            0            0
## Region                0            0            0
## GDP                   0            0            0
## IncomeStatus          0            0            0
## Population            0            0            0
## CoastalPopPC          0            0            0
## UrbanPopPC            0            0            0
## PlasticWaste          2            0            0
## MismanagedPW         41           39            0
## 
## $mm
##              Country Code Region GDP IncomeStatus Population CoastalPopPC
## Country            0    0      0   0            0          0            0
## Code               0    0      0   0            0          0            0
## Region             0    0      0   0            0          0            0
## GDP                0    0      0   0            0          0            0
## IncomeStatus       0    0      0   0            0          0            0
## Population         0    0      0   0            0          0            0
## CoastalPopPC       0    0      0   0            0          0            0
## UrbanPopPC         0    0      0   0            0          0            0
## PlasticWaste       0    0      0   0            0          0            0
## MismanagedPW       0    0      0   0            0          0            0
##              UrbanPopPC PlasticWaste MismanagedPW
## Country               0            0            0
## Code                  0            0            0
## Region                0            0            0
## GDP                   0            0            0
## IncomeStatus          0            0            0
## Population            0            0            0
## CoastalPopPC          0            0            0
## UrbanPopPC            0            0            0
## PlasticWaste          0            2            2
## MismanagedPW          0            2           41

aggr(plastic, col = mdc(1:2), numbers = TRUE,
     sortVars = TRUE, labels = names(plastic),
     cex.lab = .7,
     cex.axis = .6, gap = 2, ylab = c("Proportion of missingness","Missingness Pattern"))

## 
##  Variables sorted by number of missings: 
##      Variable       Count
##  MismanagedPW 0.194312796
##  PlasticWaste 0.009478673
##       Country 0.000000000
##          Code 0.000000000
##        Region 0.000000000
##           GDP 0.000000000
##  IncomeStatus 0.000000000
##    Population 0.000000000
##  CoastalPopPC 0.000000000
##    UrbanPopPC 0.000000000

## so, we have only two missingness pattern, with the highest proportion of missingness correspond to mismanaged PW
marginplot(plastic[, c("MismanagedPW", "PlasticWaste")],
           col = mdc(1:2),
           cex.numbers = 1.2, pch = 19, xlab = "Mismanaged Plastic Waste per capita",
           ylab = "Plastic Waste per capita")

We can check the missing cases as shown below:

x <- plastic[!complete.cases(plastic),  ] # lots of missing values for mismanaged PW

xtabs(~Region, data = x) # most of these from Sub Saharan Africa (23), East Asia & Pacific, Latin America

## Region
##        East Asia & Pacific      Europe & Central Asia 
##                          7                          2 
##  Latin America & Caribbean Middle East & North Africa 
##                          6                          2 
##              North America                 South Asia 
##                          0                          1 
##         Sub-Saharan Africa 
##                         23

xtabs(~IncomeStatus, data = x) # very high missing values from low and low-middle income (26)

## IncomeStatus
##         Low  Low-middle High-middle        High 
##          15          11           7           8

Correlation plot

A correlation plot is obtained to explore the relationships between the two types of plastic waste and other variables in the plastic dataset.

# remove missing values first to calculate correlation manually
data <- na.omit(plastic)
round(cor(data[, c(4, 6, 7, 8, 9, 10)]), 2)

##                GDP Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## GDP           1.00      -0.09         0.16       0.44         0.43        -0.05
## Population   -0.09       1.00        -0.04      -0.08        -0.11        -0.06
## CoastalPopPC  0.16      -0.04         1.00       0.15         0.02         0.03
## UrbanPopPC    0.44      -0.08         0.15       1.00         0.38         0.02
## PlasticWaste  0.43      -0.11         0.02       0.38         1.00         0.42
## MismanagedPW -0.05      -0.06         0.03       0.02         0.42         1.00

# or using corrplot package
# install.packages ('corrplot', 'RColorBrewer')

library(corrplot)

## corrplot 0.92 loaded

library(RColorBrewer)
# mat : is a matrix of data
# ... : further arguments to pass to the native R cor.test function
cor.mtest <- function(mat, ...) {
  mat <- as.matrix(mat)
  n <- ncol(mat)
  p.mat<- matrix(NA, n, n)
  diag(p.mat) <- 0
  for (i in 1:(n - 1)) {
    for (j in (i + 1):n) {
      tmp <- cor.test(mat[, i], mat[, j], ...)
      p.mat[i, j] <- p.mat[j, i] <- tmp$p.value
    }
  }
  colnames(p.mat) <- rownames(p.mat) <- colnames(mat)
  p.mat
}

# matrix of the p-value of the correlation
p.mat <- cor.mtest(data[, c(4, 6, 7, 8, 9, 10)])
head(p.mat[, 1:5])

##                       GDP Population CoastalPopPC   UrbanPopPC PlasticWaste
## GDP          0.000000e+00  0.2599869   0.03318203 1.468796e-09 3.611139e-09
## Population   2.599869e-01  0.0000000   0.64693272 2.968781e-01 1.546294e-01
## CoastalPopPC 3.318203e-02  0.6469327   0.00000000 4.679163e-02 7.620806e-01
## UrbanPopPC   1.468796e-09  0.2968781   0.04679163 0.000000e+00 4.550404e-07
## PlasticWaste 3.611139e-09  0.1546294   0.76208059 4.550404e-07 0.000000e+00
## MismanagedPW 4.959714e-01  0.4003487   0.66421522 7.999911e-01 9.764815e-09

corrplot(cor(data[, c(4, 6, 7, 8, 9, 10)]), tl.cex = 0.75, method = "color", type = "upper", order = "hclust", col = brewer.pal(n=8, name="RdYlBu"), tl.col = "red", tl.srt = 45, addCoef.col = "black",
         # Combine with significance
         p.mat = p.mat, sig.level = 0.01,
         # hide correlation coefficient on the principal diagonal
         diag=FALSE )

The correlation plot shows that Plastic waste has a relatively positive correlation with Mismanaged plastic waste (0.42), meaning countries with high plastic waste will also have high mismanaged plastic waste generated per capita. However, the correlation is only considered to be moderate since some countries particularly those with high-income status also tend to have low mismanaged plastic waste generated per capita.
Besides that, plastic waste also positively correlates with GDP per capita (0.43) and Urban population (0.38), meaning countries with high GDP per capita and high rate of urbanization tend to have high plastic waste. This is unsurprising given that a rapid economic growth can influence a high influx of migrants into urban areas, which will further increase the production of plastic materials in order to support the rising demands from populations. This finding is supported by a positive correlation between GDP per capita and the urban population features (0.44). Therefore, it is safe to conclude that countries with high-income status typically have high GDP per capita and urbanization rate, and consequently, high plastic waste per capita per day.
Apart from this, other associations do not appear to be significant and are relatively weak. For example, Mismanaged plastic waste has a very weak, negative association with GDP per capita (-0.05). This means that countries with high GDP per capita tend to have low mismanaged plastic waste. Population in general also has a weak, negative association with both types of plastic waste and other features in the data set. On the other hand, both Urban and Coastal populations have positive correlations with both types of plastic waste.

Distribution of Plastic waste

## combined plots of histogram and boxplot of plastic waste
par(mfrow = c(2, 1))
hist(plastic$PlasticWaste,
     xlab = "Plastic waste per capita (kg/day)",
     col = "skyblue",
     breaks = 25,
     main = "Histogram of plastic waste")

boxplot(plastic$PlasticWaste,
        xlab = "Plastic waste per capita (kg/day)",
        col = 'skyblue',
        horizontal = TRUE,
        main = "Boxplot of plastic waste")

## Outliers/extreme values 

# outliers for PW
out <- boxplot.stats(plastic$PlasticWaste)$out
idx <- which(plastic$PlasticWaste %in% c(out))
plastic[idx, c('Country', 'Region', 'PlasticWaste')]

##                   Country                    Region PlasticWaste
## ABW                 Aruba Latin America & Caribbean       132.53
## BMU               Bermuda             North America       205.93
## CYM        Cayman Islands Latin America & Caribbean       133.86
## FRO        Faeroe Islands     Europe & Central Asia       167.53
## HKG  Hong Kong SAR, China       East Asia & Pacific       163.45
## ISL               Iceland     Europe & Central Asia       144.80
## FSM Micronesia, Fed. Sts.       East Asia & Pacific       308.25
## MDA               Moldova     Europe & Central Asia       130.15
## MNG              Mongolia       East Asia & Pacific       137.58
## PLW                 Palau       East Asia & Pacific       170.19
## PRI           Puerto Rico Latin America & Caribbean       134.10
## SGP             Singapore       East Asia & Pacific       143.96
## KNA   St. Kitts and Nevis Latin America & Caribbean       150.01
## USA         United States             North America       130.09
## VIR Virgin Islands (U.S.) Latin America & Caribbean       206.43

The histogram and boxplot show that the distribution of Plastic waste generated per capita is very heavily skewed to the right, with extreme outliers from 125 kilograms of plastic waste per capita. A very high peak is seen at approximately 25 kilograms of plastic waste per capita with a density of 0.023. A gap between 250 kilograms and 300 kilograms of plastic waste is also seen, corresponding to two missing values in the Turks and Caicos Islands and Sint Maarten (Dutch part). It is worth noting that the data is somewhat polarizing with either low or high levels of plastic waste generated per capita, suggesting that underlying factors may influence such different levels of plastic pollution. For example, initial investigation reveals that a large cluster of low levels of plastic waste generated (i.e., 0-55 kilograms per day) correspond to Sub-Saharan Africa region with 46 observations, followed by Europe and Central Asia with 33 observations, and East Asia and Pacific with 28 observations. The other cluster corresponds to the Middle East and North Africa, South Asia and North America.
The top extreme outlier seen on the boxplot at 308.25 kilograms of plastic waste generated suggests a country with the highest level of plastic waste: the Federated States of Micronesia, a low-middle income country with a relatively small population of 110215, of which 22.53% lives in the urban area, and 8.41% lives in the coastal area. The following outliers correspond to Bermuda and the Virgin Islands, shown as two overlapping outliers at around 205-206 kilograms of plastic waste per day. Both are high-income countries with very high urban populations (i.e., 95.48% for the Virgin Islands and 100% for Bermuda) and small coastal populations (i.e., 5.44% for the Virgin Islands and 11.86% for Bermuda). Indeed, Bermuda has the highest 100% urbanization other than Singapore, China, Cayman Islands and Monaco (World Urbanization Prospects, 2018).
Notably, there are also two clusters of outliers at approximately 130-150 kilograms of plastic waste per capita and 163-170 kilograms of plastic waste per capita. These clusters all correspond to high-income countries such as the United States (130.09 kilograms per capita), Singapore (143.96 kilograms per capita), Faeroe Islands par(mfrow = c(2, 1))
(167.53 kilograms per capita), and Palau (170.19 kilograms per day) and two low-middle income countries such as Moldova (130.15 kilograms per day) and Mongolia (137.58 kilograms per day). Most of the countries within these clusters have high urbanization rates except for the aforementioned low-middle income countries and island countries such as Faeroe Islands and St. Kitts and Nevis.

Distributions of Mismanaged Plastic Waste

par(mfrow = c(2, 1))
hist(plastic$MismanagedPW,
     xlab = "Mismanaged plastic waste per capita (kg/day)",
     breaks = 25,
     col = "thistle",
     main = "Histogram of mismanaged plastic")
boxplot(plastic$MismanagedPW,
        xlab = "Mismanaged plastic waste per capita (kg/day)",
        col = "thistle",
        horizontal = TRUE,
        main = "Boxplot of mismanaged plastic")

# outliers for mismanaged PW
out <- boxplot.stats(plastic$MismanagedPW)$out
idx <- which(plastic$MismanagedPW %in% c(out))
plastic[idx, c('Country', 'Region', 'MismanagedPW')]

##                      Country                     Region MismanagedPW
## ABW                    Aruba  Latin America & Caribbean       120.60
## CYM           Cayman Islands  Latin America & Caribbean       108.43
## FRO           Faeroe Islands      Europe & Central Asia        58.64
## KAZ               Kazakhstan      Europe & Central Asia        71.36
## KWT                   Kuwait Middle East & North Africa       115.77
## MDA                  Moldova      Europe & Central Asia       143.00
## MNG                 Mongolia        East Asia & Pacific       137.58
## MNE               Montenegro      Europe & Central Asia        97.07
## MNP Northern Mariana Islands        East Asia & Pacific        59.93
## OMN                     Oman Middle East & North Africa        90.44
## RUS       Russian Federation      Europe & Central Asia        57.19
## SLB          Solomon Islands        East Asia & Pacific        69.34
## TTO      Trinidad and Tobago  Latin America & Caribbean        90.77
## TUV                   Tuvalu        East Asia & Pacific        51.01
## ARE     United Arab Emirates Middle East & North Africa        70.32

The histogram and boxplot show that the distribution of Mismanaged plastic waste per capita per day is very heavily skewed to the right, with extreme outliers from around 45 kilograms and above. A very high peak is seen around 10 kilograms with an approximate density of 0.073 - about three times higher than the density peak for plastic waste. Interestingly, this peak corresponds to high-middle- and high-income countries, which account for the highest proportions of the lowest mismanaged plastic waste. This initial observation may suggest that high-income countries have a good waste management system that influences a low level of mismanaged plastic waste, despite having a high level of plastic waste.
The boxplot shows a few clusters of outliers, corresponding to a mix of high, high-middle and low-middle income countries. For example, the two top extreme outliers correspond to two low-middle income countries: Mongolia (137.58 kilograms per capita) and Moldova (143 kilograms per capita); both are also extreme outliers for plastic waste. The following outliers correspond to high-income countries with a 100% urbanization rate, such as the Cayman Islands (108.43 kilograms per capita) and Kuwait (115.77 kilograms per capita) and Aruba (120.60 kilograms per day). This trend suggests that urbanization is also likely to affect mismanaged plastic waste per capita per day.
Surprisingly, low-income countries such as Guinea (3.11 kilograms per capita), Ethiopia (1.49 kilograms per capita), Nigeria (4.04 kilograms per capita) and other low-middle income countries such as Bangladesh (4.69 kilograms per capita), Bhutan (0.54 kilograms per capita), and Cameroon (4.11 kilograms per capita) have a relatively low level of mismanaged plastic waste. It is also possible that these values may not reflect the accurate picture due to a high proportion of data missingness in low-income countries (i.e., 63.4% of the total missingness). Overall, initial explorations of the distributions of these two types of plastic waste suggest that different regions, countries’ income status levels, and urbanization likely affect plastic pollution.

Distributions of plastic waste across regions

# install.packages('lattice', 'ggplot2', 'ggridges', 'ggpubr')
# Plastic waste split by regions
library(lattice)
histogram(~PlasticWaste|Region, data = plastic,
          xlab = "Plastic Waste (kg/day)")

library(ggplot2)
library(ggridges)
library(ggpubr)

# ridge plot, boxplot, histograms by regions

# ridge plot
ggplot(plastic,aes(PlasticWaste, y = factor(Region),fill = Region))+
  geom_density_ridges(scale = 1.5) +
  labs(x = "Plastic waste per capita (kg/day)",
       y = "Region") + 
  ggtitle("Plastic waste by regions") +
    theme(plot.title = element_text(hjust = 0.5, face = "bold"))

## Picking joint bandwidth of 16.9

## Warning: Removed 2 rows containing non-finite values (stat_density_ridges).

# boxplot
ggplot(plastic,aes(x = PlasticWaste, y = factor(Region))) +
geom_boxplot(aes(fill = factor(Region))) +
labs(y = "Region", x = "Plastic waste per capita (kg/day)")

## Warning: Removed 2 rows containing non-finite values (stat_boxplot).

# outliers
# outliers in region 2

out <- boxplot.stats(plastic$PlasticWaste[plastic$Region == "Europe & Central Asia"])$out
idx <- which(plastic$PlasticWaste %in% c(out))
plastic[idx,]

##            Country Code                Region     GDP IncomeStatus Population
## FRO Faeroe Islands  FRO Europe & Central Asia 56854.1         High      48173
## ISL        Iceland  ISL Europe & Central Asia 61466.8         High     335439
##     CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## FRO         6.77      41.78       167.53        58.64
## ISL         7.56      93.74       144.80        16.13

# outliers in region 3
out <- boxplot.stats(plastic$PlasticWaste[plastic$Region == "Latin America & Caribbean"])$out
idx <- which(plastic$PlasticWaste %in% c(out))
plastic[idx,]

##                   Country Code                    Region      GDP IncomeStatus
## KNA   St. Kitts and Nevis  KNA Latin America & Caribbean 18810.49         High
## VIR Virgin Islands (U.S.)  VIR Latin America & Caribbean 35931.54         High
##     Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## KNA      51629         3.89      30.79       150.01            3
## VIR     107510         5.44      95.48       206.43           NA

# outliers in region 7
out <- boxplot.stats(plastic$PlasticWaste[plastic$Region == "Sub-Saharan Africa"])$out
idx <- which(plastic$PlasticWaste %in% c(out))
plastic[idx,]

##        Country Code             Region       GDP IncomeStatus Population
## MUS  Mauritius  MUS Sub-Saharan Africa  9681.619  High-middle    1263473
## SYC Seychelles  SYC Sub-Saharan Africa 15068.620         High      94677
##     CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## MUS         1.41      40.91        42.94         0.86
## SYC        16.64      55.83        56.38           NA

#outliers in region 6
out <- boxplot.stats(plastic$PlasticWaste[plastic$Region == "South Asia"])$out
idx <- which(plastic$PlasticWaste %in% c(out))
plastic[idx,]

##  [1] Country      Code         Region       GDP          IncomeStatus
##  [6] Population   CoastalPopPC UrbanPopPC   PlasticWaste MismanagedPW
## <0 rows> (or 0-length row.names)

The plots above show the distributions of plastic waste generated per capita across different regions. The ridgeline plot shows that Sub-Saharan Africa and South Asia have a relatively similar peak at around 12.50 kilograms of plastic waste per capita. Thus, surprisingly, marking these two regions as having the lowest level of plastic waste generated per capita per day compared to the other regions.The density distribution for South Asia appears to be relatively normal, while the density distribution for Sub- Saharan Africa seems to be right-skewed.
On the other hand, North America has a trimodal distribution with three peaks observed at three different levels of plastic waste generated per capita. Initial exploration reveals that these peaks correspond to three countries within this region: Canada (21.32 kilograms per capita), the United States (130.09 kilograms per capita) and Bermuda (205.93 kilograms per capita). Thus, this region has the largest volume of plastic waste generated, with a uniform distribution since all three peaks have almost the same height. The rest of the regions all have a relatively right-skewed distribution with a very similar peak between 42-50 kilograms of plastic waste generated per capita.
In addition, there are several outliers observed in most of the regions except for North America and the Middle East and North Africa. For example, the outliers in the East Asia and Pacific region correspond to Hong Kong, China (163.45 kilograms per capita), Palau (170.19 kilograms per capita) and Micronesia (308.25 kilograms per capita); in the Europe and Central Asia region, the outliers correspond to Iceland (144.80 kilograms per capita) and Faeroe Islands (144.80 kilograms per capita), in the Latin America and Caribbean region the outliers correspond to St. Kitts and Nevis (150.01 kilograms per capita) and Virgin Islands (206.43 kilograms per capita), in the Sub- Saharan Africa region the outliers correspond to Mauritius (42.94 kilograms per capita) and Seychelles (56.38 kilograms per capita). In South Asia, the outliers correspond to Bhutan (26.94 kilograms per capita).

Distributions of Mismanaged plastic waste across regions

# lattice
histogram(~MismanagedPW|Region, data = plastic,
          xlab = "Mismanaged Plastic Waste (kg/day)")

# ridge plot
ggplot(plastic,aes(MismanagedPW, y = factor(Region),fill = Region))+
  geom_density_ridges(scale = 1.5) +
  labs(x = "Mismanaged plastic waste per capita (kg/day)",
       y = "Region") + 
  ggtitle("Mismanaged plastic waste by regions") +
  theme(plot.title = element_text(hjust = 0.5, face = "bold"))

## Picking joint bandwidth of 4.47

## Warning: Removed 41 rows containing non-finite values (stat_density_ridges).

# boxplot
ggplot(plastic,aes(x = MismanagedPW, y = factor(Region))) +
  geom_boxplot(aes(fill = factor(Region))) +
  labs(y = "Region", x = "Mismanaged plastic waste per capita (kg/day)")

## Warning: Removed 41 rows containing non-finite values (stat_boxplot).

# outliers in region 1

out <- boxplot.stats(plastic$MismanagedPW[plastic$Region == "East Asia & Pacific"])$out
idx <- which(plastic$MismanagedPW %in% c(out))
plastic[idx,]

##                      Country Code              Region       GDP IncomeStatus
## MNG                 Mongolia  MNG East Asia & Pacific  3660.151   Low-middle
## MNP Northern Mariana Islands  MNP East Asia & Pacific 22246.743         High
## SLB          Solomon Islands  SLB East Asia & Pacific  2225.490   Low-middle
## TUV                   Tuvalu  TUV East Asia & Pacific  3255.929  High-middle
##     Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## MNG    3056359         0.00      68.30       137.58       137.58
## MNP      56188         1.85      91.44        90.80        59.93
## SLB     619437         2.86      22.82        69.34        69.34
## TUV      11225        47.47      60.65        58.64        51.01

# outliers in region 2

out <- boxplot.stats(plastic$MismanagedPW[plastic$Region == "Europe & Central Asia"])$out
idx <- which(plastic$MismanagedPW%in% c(out))
plastic[idx,]

##                Country Code                Region       GDP IncomeStatus
## FRO     Faeroe Islands  FRO Europe & Central Asia 56854.103         High
## KAZ         Kazakhstan  KAZ Europe & Central Asia  7714.842  High-middle
## MDA            Moldova  MDA Europe & Central Asia  2880.439   Low-middle
## MNE         Montenegro  MNE Europe & Central Asia  7028.663  High-middle
## RUS Russian Federation  RUS Europe & Central Asia  8704.898  High-middle
##     Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## FRO      48173         6.77      41.78       167.53        58.64
## KAZ   17794055         0.00      57.26        72.02        71.36
## MDA    2802170         0.59      42.51       130.15       143.00
## MNE     622303         0.00      66.14       100.50        97.07
## RUS  144342396         1.21      74.16        58.66        57.19

# outliers in region 3
out <- boxplot.stats(plastic$MismanagedPW[plastic$Region == "Latin America & Caribbean"])$out
idx <- which(plastic$MismanagedPW %in% c(out))
plastic[idx,]

##                 Country Code                    Region      GDP IncomeStatus
## ABW               Aruba  ABW Latin America & Caribbean 28281.35         High
## CYM      Cayman Islands  CYM Latin America & Caribbean 78465.36         High
## TTO Trinidad and Tobago  TTO Latin America & Caribbean 16176.95         High
##     Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## ABW     104872         7.07      43.19       132.53       120.60
## CYM      62569        12.07     100.00       133.86       108.43
## TTO    1377564         2.91      53.25       101.75        90.77

# outliers in region 4
out <- boxplot.stats(plastic$MismanagedPW[plastic$Region == "Middle East & North Africa"])$out
idx <- which(plastic$MismanagedPW %in% c(out))
plastic[idx,]

##                  Country Code                     Region      GDP IncomeStatus
## KWT               Kuwait  KWT Middle East & North Africa 27653.07         High
## OMN                 Oman  OMN Middle East & North Africa 14618.73         High
## ARE United Arab Emirates  ARE Middle East & North Africa 38141.85         High
##     Population CoastalPopPC UrbanPopPC PlasticWaste MismanagedPW
## KWT    3956875        17.40     100.00       115.77       115.77
## OMN    4479219         5.65      82.50        90.44        90.44
## ARE    9360980        14.78      85.97       109.88        70.32

The plots above show the distributions of Mismanaged plastic waste generated per capita across regions. The distribution appears to be primarily right skewed, except for North America, with a relatively uniform distribution. South Asia, the Middle East, and North Africa have a relatively normal distribution. It turns out that North America has the lowest level of mismanaged plastic waste generated per capita with an average of 2.69 kilograms, despite having the highest level of plastic waste generated per capita, as shown previously in Figure 4. Within this region, Canada accounts for the lowest value of the mismanaged plastic waste generated (0.43 kilograms per capita), followed by the United States (3.51 kilograms per capita) and Bermuda (4.12 kilograms per capita).
Similarly, Europe and Central Asia have a similar distribution peak to North America, possibly indicating that this region has good waste management infrastructures and systems that control the amount of plastic waste inadequately discarded in the surrounding environment. However, it is worth noting that its density distribution appears to be relatively right skewed as some other countries within this region have a high level of mismanaged plastic waste generated per capita. For example, Moldova (143.00 kilograms per capita), Kazakhstan (71.36 kilograms per capita) and Russia (97.07 kilograms per capita) as indicated by the outliers.
The Middle East and North Africa appear to have the highest level of mismanaged plastic waste generated per capita per day (5b), with an average of 24.63 kilograms. The density distribution appears to be normally distributed. However, there are three small peaks within this region corresponding to three outliers, which are identified as Oman (90.44 kilograms per capita), United Arab Emirates (109.88 kilograms per capita) and Kuwait (115.77 kilograms per capita). These three countries turn out to be high-income countries.
Besides that, Latin America and the Caribbean has a similar average value of mismanaged plastic waste to the Middle East and North Africa, but with a very right- skewed distribution. The outliers within this region correspond to Trinidad and Tobago (90.77 kilograms per capita), Cayman Islands (108.43 kilograms per capita), and Aruba (132.53 kilograms per capita).
South Asia and Sub-Saharan Africa also have a similar distribution peak at around 11.22 kilograms of mismanaged plastic waste generated per capita per day, which is relatively low. There are no outliers observed within these two regions. Meanwhile, in East Asia and the Pacific, there are several outliers observed corresponding to Tuvalu (51.01 kilograms per capita), Northern Marian Islands (59.93 kilograms per capita), Solomon Islands (69.34 kilograms per capita), and Mongolia (137.58 kilograms per capita). Overall, we see quite a difference between the mismanaged plastic waste generated and the plastic waste generated per capita across different regions. Perhaps, other features such as income status may provide further interesting explanation for this difference.

Distributions of plastic waste across income

# Plastic waste split by income status
histogram(~PlasticWaste|IncomeStatus, data = plastic,
          xlab = "Plastic Waste (kg/day)")

# violin plot
ggplot(plastic,
       aes(x = factor(IncomeStatus),
           y = PlasticWaste)) +
  geom_violin(aes(fill = IncomeStatus)) +
  labs(x = "Income Status", y = "Plastic waste per capita (kg/day)")+
  geom_boxplot(width = 0.08) + scale_fill_discrete(name = "Income Status") + 
  ggtitle("Plastic waste by income status") +
  theme(plot.title = element_text(face = "bold", hjust = 0.5),
        legend.position = "bottom")

## Warning: Removed 2 rows containing non-finite values (stat_ydensity).

## Warning: Removed 2 rows containing non-finite values (stat_boxplot).

# histogram

ggplot(plastic,
      aes(PlasticWaste, fill = IncomeStatus
          )) +
  geom_histogram(binwidth = 5,
                 aes(y = ..density..),
                 colour = 1, fill = "white") + labs( x = "Plastic waste per capita (kg/day)")+
  facet_wrap(~IncomeStatus) + 
  geom_density(lwd = 0.4, colour = 4, fill = 4, alpha = 0.24)

## Warning: Removed 2 rows containing non-finite values (stat_bin).

## Warning: Removed 2 rows containing non-finite values (stat_density).

# ridge plot
ggplot(plastic,aes(PlasticWaste, y = factor(IncomeStatus),fill = IncomeStatus))+
  geom_density_ridges(scale = 1.5) +
  labs(x = "Plastic waste per capita (kg/day)",
       y = "Income Status") + 
  ggtitle("Ridgeline plot") +
  theme(plot.title = element_text(hjust = 0.5, face = "bold"),
        legend.position = "none")

## Picking joint bandwidth of 8.15

## Warning: Removed 2 rows containing non-finite values (stat_density_ridges).

The plots above show the distribution of plastic waste generated per capita across countries of different income statuses. The boxplots inside the density curves suggest that high-income countries have the highest number of plastic wastes generated, followed by high-middle, low- middle, and low-income countries. There are also outliers observed within each category. For example, there are three overlapping outliers in the low-income category corresponding to Haiti (27.06 kilograms per capita), Rwanda (28.18 kilograms per capita) and Zimbabwe (29.49 kilograms per capita). For low-middle income countries, the outliers correspond to Solomon Islands (69.34 kilograms per capita), Moldova (130.15 kilograms per capita), Mongolia (137.58 kilograms per capita) and Micronesia (308.25 kilograms per capita). For high-middle income countries, the outliers correspond to St. Lucia (95.76 kilograms per capita) and Montenegro (100.50 kilograms per capita). Lastly, for high-income countries, the outliers correspond to Bermuda (205.93 kilograms per capita) and the Virgin Islands (206.43 kilograms per capita).
In terms of the overall density distribution, the low and high-middle income countries appear to be normally distributed with distinct peaks at around 12.00 kilograms per capita and 36.00 kilograms per capita, respectively. On the other hand, the low-middle- and high-income countries have a rather right-skewed distribution, with the latter having a much-elongated distribution with a lack of distinct peak. Nevertheless, there is a clear pattern of association between the income status and the plastic waste generated per capita, which may explain the previously seen differences across varying regions.

Distributions of mismanaged plastic waste across income

# Mimanaged PW split by income status
histogram(~MismanagedPW|IncomeStatus, data = plastic,
          xlab = "Mismanaged Plastic Waste (kg/day)")

# violin plot
ggplot(plastic,
                  aes(x = factor(IncomeStatus),
                      y = MismanagedPW)) +
  geom_violin(aes(fill = IncomeStatus)) +
  labs(x = "Income Status", y = "Mismanaged plastic waste per capita (kg/day)")+
  geom_boxplot(width = 0.08) + scale_fill_discrete(name = "Income Status") + 
  ggtitle("Mismanaged plastic waste by income status") +
  theme(plot.title = element_text(face = "bold", hjust = 0.5),
        legend.position = "bottom")

## Warning: Removed 41 rows containing non-finite values (stat_ydensity).

## Warning: Removed 41 rows containing non-finite values (stat_boxplot).

# ridge plot
ggplot(plastic,
       aes(MismanagedPW, y = factor(IncomeStatus),
           fill = IncomeStatus))+
  geom_density_ridges(scale = 1.5) +
  labs(x = "Mismanaged Plastic Waste (kg/day)",
       y = "Income Status")

## Picking joint bandwidth of 3.99

## Warning: Removed 41 rows containing non-finite values (stat_density_ridges).

The plots above show the distribution of mismanaged plastic waste generated per capita across countries with different income statuses. In contrast to the previous finding, high-income countries have the lowest mismanaged plastic waste generated, followed by the low-middle, high-middle, and low-income countries. This trend suggests that high-income countries have good waste management systems and infrastructures to properly control the plastic waste from being discarded into the surrounding environment.
Most distributions appear to be relatively right-skewed with distinct peaks, except for the low-income countries with a rather left-skewed distribution. Several outliers are identified, especially in high and high-middle income countries, indicating that not all countries with the same income status have the same waste management systems and infrastructures. For example, Aruba (120.60 kilograms per capita), Kuwait (115.00 kilograms per capita), and Cayman Islands (108.43 kilograms per capita) are the topmost extreme outliers in the high-income category. These countries also have high plastic waste generated. Nevertheless, it does appear that a high-income status positively contributes to the low level of mismanaged plastic waste generated per capita per day. Still, it also negatively contributes to the high level of plastic waste generated per capita per day.

Relationships between other variables and plastic waste

Parallel coordinates plot can be used to show relationships between variables. The plot below shows the relationships between variables and different income status is highlighted in distinct colours for comparison.
Indeed, the purple lines corresponding to high-income countries tend to form the top layers of observations on each feature, meaning that those countries tend to dominate in terms of the GDP per capita, urbanization and plastic waste.
On the other hand, the blue and green lines corresponding to the low-middle and high-middle income countries, respectively, tend to follow a similar pattern to those seen in high-income countries. However, the only difference is that the mismanaged plastic waste tends to be higher than the plastic waste generated per capita.
In addition, there are two striking blue and green points on the population axis, which turn out to be two outliers corresponding to a high-middle income country, China, and a low-middle income country, India, with a very high number of populations. Lastly, the low-income countries represented as the orange lines tend to form the bottommost layer of the parallel coordinates plot, meaning these countries have low values on these features.

library(MASS)
library(GGally)

## Registered S3 method overwritten by 'GGally':
##   method from   
##   +.gg   ggplot2

## vector colour
library(RColorBrewer)
palette <- brewer.pal(4, "Set1")
my_colors <- palette[as.numeric(plastic$IncomeStatus)]

levels(plastic$IncomeStatus) <- c("Low", "Low-middle", "High-middle", "High")
ggparcoord(plastic,
           columns = c(4, 6, 7, 8, 9, 10),
           groupColumn = 5,
           showPoints = TRUE,
           scale = "uniminmax",
           missing = "exclude",
           title = "Parallel coordinates of plastic data by income group",
           alphaLines = 0.4) +
  theme(
    plot.title = element_text(size=10, face = "bold"),
    legend.position = "bottom") +scale_shape_discrete(name = "Income")

GDP and mismanaged/plastic waste

The interaction between GDP and the two types of plastic waste entails an interesting feature.

lfit <- loess(PlasticWaste ~ GDP, data=plastic, span=0.66)
## create a sequence of points to cover the x-axis
xs <- seq(min(plastic$GDP), max(plastic$GDP), length=200)
## get the predicted trend at the new locations, and the error too
lpred <- predict(lfit, data.frame(GDP=xs), se = TRUE) 
plot(plastic$GDP,
     plastic$PlasticWaste,
     xlab = "GDP per capita",
     ylab =  "Plastic waste per capita (kg/day)",
     main = "Plastic waste vs GDP",
     pch =16)
## add the loess prediction
lines(x=xs, y=lpred$fit, col='red',lwd=4) 

##  loess curve is much smoother and better behaved, however it is still influenced by the outliers in the data

## EXTRACT CI
upr <- lpred$fit + 1.96*lpred$se.fit
lwr <- lpred$fit - 1.96*lpred$se.fit
lines(x=xs, y=lpred$fit,col='red',lwd=4)
lines(x=xs, y=upr,col='red',lwd=1, lty=2)
lines(x=xs, y=lwr,col='red',lwd=1, lty=2)

## These confidence intervals are intervals for the location of the trend, not for the data. We do not expect most of our data to be between these limits.
## Notice how our uncertainty grows to the right of the plot where we have fewest data.

## USE POLYGONS

## we need 'scales' to use the alpha function for transparency
library(scales) 
polygon(x=c(xs,rev(xs)), 
        y=c(upr,rev(lwr)),
        col=alpha('red',0.2), border=NA)

# mismanaged pw vs gdp -----


lfit <- loess(MismanagedPW ~ GDP, data=plastic, span=0.66)
## create a sequence of points to cover the x-axis
xs <- seq(min(plastic$GDP), max(plastic$GDP), length=200)
## get the predicted trend at the new locations, and the error too
lpred <- predict(lfit, data.frame(GDP=xs), se = TRUE)
plot(plastic$GDP,
     plastic$MismanagedPW,
     xlab = "GDP per capita",
     ylab =  "Mismanaged plastic waste per capita (kg/day)",
     main = "Mismanaged plastic waste vs GDP",
     pch =16)

## add the loess prediction
lines(x=xs, y=lpred$fit, col='red',lwd=4) 

##  loess curve is much smoother and better behaved, however it is still influenced by the outliers in the data

## EXTRACT CI
upr <- lpred$fit + 1.96*lpred$se.fit
lwr <- lpred$fit - 1.96*lpred$se.fit
lines(x=xs, y=lpred$fit,col='red',lwd=4)
lines(x=xs, y=upr,col='red',lwd=1, lty=2)
lines(x=xs, y=lwr,col='red',lwd=1, lty=2)

## These confidence intervals are intervals for the location of the trend, not for the data. We do not expect most of our data to be between these limits.
## Notice how our uncertainty grows to the right of the plot where we have fewest data.

## USE POLYGONS

## we need 'scales' to use the alpha function for transparency
library(scales) 
polygon(x=c(xs,rev(xs)), 
        y=c(upr,rev(lwr)),
        col=alpha('red',0.2), border=NA)



## FIT LINEAR 
lmfit <- lm(MismanagedPW ~ GDP, 
            data = plastic)
abline(lmfit, col = "green")

Plotting these plastic wastes against the GDP per capita suggests that plastic waste tends to increase as GDP per capita increases. Still, mismanaged plastic waste tends to decrease. Notice that the uncertainty grows to the right of the plot, where the data are scarce. The local regression curves appear to be relatively smooth; however, they are still greatly influenced by extreme outliers in the data and heavily right-skewed GDP data. Nevertheless, it is safe to link the GDPs per capita and the two types of plastic waste.

Bubble plot

Bubble plots are also obtained to explore the relationships between income status, regions, and population on the two types of plastic waste. The plot below shows that there is a pretty clear separation on the level of plastic waste generated per capita between different regions: Sub-Saharan Africa appears to form a cluster in the bottom left corner of the plot, indicating that this region has the lowest GDP per capita and produces low plastic waste per capita. The bubbles’ size for this region appears to be relatively small compared to others, indicating a small population size associated with this region. As we move along to the right of the plot, South Asia (i.e., purple bubbles) appears to be next to Sub- Saharan Africa.
Within South Asia, India seems to have the largest population size and a relatively high plastic waste generated per capita (19.88 kilograms per capita) – the second highest after Bhutan (26.94 kilograms per capita).This trend is followed by East Asia and the Pacific (i.e., red bubbles), in which mainland China appears to have the largest population size. Interestingly, China’s plastic waste per capita is distinctively lower than any other smaller country within this region (i.e., only 15.67 kilograms per capita compared to Micronesia (308.25 kilograms per capita)).
With a much lower population, Hong Kong has a much higher plastic waste than in China (i.e., the top right red bubble on top of the US, representing 163.45 kilograms of plastic waste per capita). Next, we have Latin America and the Caribbean (i.e., pale green bubbles) and the Middle East and North Africa (i.e., green bubbles), followed by a much higher GDP and plastic waste generated per capita in Europe and Central Asia (i.e., yellow bubbles) and lastly, North America (i.e., blue bubbles). Notice that the United Kingdom in Europe and the United States in the North America region have a comparably very high GDP per capita, and the plastic waste generated per capita, compared to other countries within the same region and in other areas.

Plastic waste against GDP across regions

par(mfrow = c(1, 1))
plot(x=log10(plastic$GDP),
     y=plastic$PlasticWaste,
     col=rainbow(n = 7, alpha = 0.8),
     xlab='GDP', ylab='Plastic Waste (kg/day)', pch=16
     , main = "Plastic waste against GDP across regions")

par(mfrow = c(1, 1))
symbols(y = plastic$PlasticWaste,
        x = log(plastic$GDP),
        bg = rainbow(n = 7, alpha = 0.8)[plastic$Region],
        fg = NA,
        circles = sqrt(plastic$Population),
        inches = 0.5,
        xlab = "log(GDP)",
        ylab = "Plastic Waste (kg/day)",
        xlim = c(2.5, 13),
        ylim = c(0, 350),
        main ="Plastic waste against GDP across regions" 
      
      )

legend(x = 'topleft',
       col = rainbow(n = 7, alpha = 0.9),
       pch = 16,
       bty = "n",
       legend = levels(plastic$Region))


# add labels CH, india, us, uk
text(x=log(plastic$GDP[plastic$Country=="China"]),
     y=plastic$PlasticWaste[plastic$Country=="China"],
     labels='CHINA', cex=0.8)

text(x=log(plastic$GDP[plastic$Country=="United States"]),
     y=plastic$PlasticWaste[plastic$Country=="United States"],
     labels='US', cex=0.8)
# UK
text(x=log(plastic$GDP[plastic$Country=="United Kingdom"]),
     y=plastic$PlasticWaste[plastic$Country=="United Kingdom"],
     labels='UK', cex=0.8)

# India
text(x=log(plastic$GDP[plastic$Country=="India"]),
     y=plastic$PlasticWaste[plastic$Country=="India"],
     labels='INDIA', cex=0.8)

text(x=log(plastic$GDP[plastic$Country=="Hong Kong SAR, China"]),
     y=plastic$PlasticWaste[plastic$Country=="Hong Kong SAR, China"],
     labels='HK', cex=0.8)

Mismanaged PW against GDP across regions

symbols(y = plastic$MismanagedPW,
        x = log(plastic$GDP),
        bg = rainbow(n = 7, alpha = 0.75)[plastic$Region],
        fg = NA,
        circles = sqrt(plastic$Population),
        inches = 0.5,
        xlab = "(log) GDP",
        ylab = "Mismanaged Waste (kg/day)",
        main = "Mismanaged plastic waste against GDP across regions",
        xlim = c(2.5, 13),
        ylim = c(0, 145)
        
)

legend(x = 'topleft',
       col = rainbow(n = 7, alpha = 0.75),
       pch = 16,
       bty = "n",
       legend = levels(plastic$Region))
# add labels CH, india, us, uk
text(x=log(plastic$GDP[plastic$Country=="China"]),
     y=plastic$MismanagedPW[plastic$Country=="China"],
     labels='CHINA', cex=0.8)

text(x=log(plastic$GDP[plastic$Country=="United States"]),
     y=plastic$MismanagedPW[plastic$Country=="United States"],
     labels='US', cex=0.8)
# UK
text(x=log(plastic$GDP[plastic$Country=="United Kingdom"]),
     y=plastic$MismanagedPW[plastic$Country=="United Kingdom"],
     labels='UK', cex=0.8)

# India
text(x=log(plastic$GDP[plastic$Country=="India"]),
     y=plastic$MismanagedPW[plastic$Country=="India"],
     labels='INDIA', cex=0.8)

Plastic waste against GDP across income status

Plotting a similar bubble plot but with grouped income status reveals a pattern on the difference in the plastic waste generated: low- income countries (i.e., red bubbles), which are identified to be in Sub-Saharan Africa, tend to sit near the bottom of the plot due to low GDP per capita and plastic waste per capita. This is followed by low-middle income countries (i.e., green bubbles) such as India, high-middle income countries (i.e., blue bubbles) such as China, and high-income countries (i.e., purple bubbles) such as the United Kingdom and the United States.

symbols(y = plastic$PlasticWaste,
        x = log(plastic$GDP),
        bg = rainbow(n = 4, alpha = 0.8)[plastic$IncomeStatus],
        fg = NA,
        circles = sqrt(plastic$Population),
        inches = 0.5,
        xlab = "log(GDP)",
        ylab = "Plastic Waste (kg/day)",
        xlim = c(2.5, 13),
        ylim = c(0, 350),
        main ="Plastic waste against GDP across income status" 
        
)
levels(plastic$IncomeStatus) <- c('Low-income', "Low-middle income", "High-middle income", "High income")
legend(x = 'topleft',
       col = rainbow(n = 4, alpha = 0.9),
       pch = 16,
       bty = "n",
       legend = levels(plastic$IncomeStatus))

# add labels CH, india, us, uk
text(x=log(plastic$GDP[plastic$Country=="China"]),
     y=plastic$PlasticWaste[plastic$Country=="China"],
     labels='CHINA', cex=0.8)

text(x=log(plastic$GDP[plastic$Country=="United States"]),
     y=plastic$PlasticWaste[plastic$Country=="United States"],
     labels='US', cex=0.8)
# UK
text(x=log(plastic$GDP[plastic$Country=="United Kingdom"]),
     y=plastic$PlasticWaste[plastic$Country=="United Kingdom"],
     labels='UK', cex=0.8)

# India
text(x=log(plastic$GDP[plastic$Country=="India"]),
     y=plastic$PlasticWaste[plastic$Country=="India"],
     labels='INDIA', cex=0.8)

Mismanaged plastic waste against GDP across income status

symbols(y = plastic$MismanagedPW,
        x = log(plastic$GDP),
        bg = rainbow(n = 4, alpha = 0.75)[plastic$IncomeStatus],
        fg = NA,
        circles = sqrt(plastic$Population),
        inches = 0.5,
        xlab = "(log) GDP",
        ylab = "Mismanaged Waste (kg/day)",
        main = "Mismanaged plastic waste against GDP across income status",
        xlim = c(2.5, 13),
        ylim = c(0, 145)
        
)

legend(x = 'topleft',
       col = rainbow(n = 4, alpha = 0.75),
       pch = 16,
       bty = "n",
       legend = levels(plastic$IncomeStatus))
# add labels CH, india, us, uk
text(x=log(plastic$GDP[plastic$Country=="China"]),
     y=plastic$MismanagedPW[plastic$Country=="China"],
     labels='CHINA', cex=0.8)

text(x=log(plastic$GDP[plastic$Country=="United States"]),
     y=plastic$MismanagedPW[plastic$Country=="United States"],
     labels='US', cex=0.8)
# UK
text(x=log(plastic$GDP[plastic$Country=="United Kingdom"]),
     y=plastic$MismanagedPW[plastic$Country=="United Kingdom"],
     labels='UK', cex=0.8)

# India
text(x=log(plastic$GDP[plastic$Country=="India"]),
     y=plastic$MismanagedPW[plastic$Country=="India"],
     labels='INDIA', cex=0.8)

For the mismanaged plastic waste, the plot below shows a trend that high-income countries with a very high GDP per capita, such as the United States in North America region and the United Kingdom in the Europe region, have a very low mismanaged plastic waste generated per capita per day. Certain high-middle income countries such as China in the East Asia and Pacific region appear to have comparably well waste management systems and infrastructures.
Low-middle-income countries like India appear to have a relatively high mismanaged plastic waste per capita. Low- income countries in Sub-Saharan Africa have low values; however, it is noted that there are high missing values for the mismanaged plastic waste data in this region, which may influence the result.

Conclusion

Regions of high-income countries with high GDP per capita tend to have higher plastic waste generated than the lower-income incomes with low GDP per capita
However, these regions tend to have lower mismanaged plastic waste that their counterparts potentially due to a better waste management system.
High urbanization further influences high plastic waste and low mismanaged plastic waste potentially due to an increased number of population living in the urban areas and better infrastructures and systems for waste management.

Plastic Pollution

Zarina

24/05/2022